1. Field of the Invention
This invention relates to temperature-sensing devices, specifically to an improved apparatus for monitoring temperature that integrates a thermistor construct into a connector construct.
2. Description of the Related Art
Temperature-sensing devices have previously been manufactured as two general types of goods. Discrete electronic thermistor components are mounted on circuit boards, or attached in wire runs. Monitoring probes, as another class, are attached on devices or placed in proximity to them, as might be typified in laboratory test environments.
Both classes of temperature-sensing devices are characterized by a thermally-resistive element, such as a thin section of metal that changes its resistance at it warms or cools. The temperature-reactive element is traditionally affixed to two or more conductors. In the case of board-mountable thermistor components, the conductors are solderable lands. Probe-type units attach to other devices with 2-4 lengths of wire. The type of conductors employed limit the way these temperature-sensors can be used in a circuit, especially where they can be located and how they can be installed. For example, if an already-manufactured power source's internal temperatures are to be monitored, board-mountable components are not a feasible solution without opening up the power source's enclosure. Even if there is access to the internal areas of the power source, for example a rechargeable battery pack, the component-style thermistor would require the correctly-configured attachment points on a circuit board.
A probe-style device offers more flexibility as to methods of attachment, for its leads can tie into an internal circuit board.
But, what options are available if there is no internal circuit board? For power sources like battery packs, for example, some have "smart" internal circuits, while others do not. If there is no internal circuitry, neither type of thermistor affords easy implementation of temperature sensing. A probe-style sensor could be installed, assuming that the battery pack could be opened (most are sealed). But the probe's wires would have to pass through a hole made in the housing, which is not very practical. Even if an access point could be created, the thick wires routed along the outside of the battery pack would likely make it difficult to reinstall the modified battery pack into a host device's molded battery cavity. Furthermore, such a radical modification to a manufactured battery pack would require significant skills, and the modified device would be aesthetically compromised.
Beside the physical limitations imposed by the traditional form factors of existing temperature-sensors, these device's modes of operation and electromechanical characteristics create further limitations.
Flexible Circuit Boards
"Flex" circuit boards are commonplace in today's small and lightweight electronic devices. Phillips Electronics (Sunnyvale, Calif.) incorporates a small flexible circuit in its two-cell rechargeable battery pack used in that company's Velo handheld computer. A discrete thermistor component is mounted on the Velo's flex-board. In particular, Duracell (Bethell, Conn.) used a component-style thermistor on a flexible circuit in its now discontinued "smart" batteries. Neither of these devices is upgradeable by the addition of a second temperature sensor, even though the Duracell's internal flex-board had a provision for a second temperature sensor. While a probe-style ancillary thermistor could be attached to either the Duracell or Phillips battery packs, as a post-manufacture retrofit, none of the various discrete thermistor types discussed below would provide its own data interface.
Thus, while flex-circuits do routinely incorporate component-style thermistors, none of these data and power circuits incorporates an integral thin and flexible thermistor as one of the traces on the circuit board itself. They therefore lack temperature sensing capabilities, combined with data and power interfaces appropriate for use on already-manufactured power sources, such as "smart batteries," for example.
As a result of this absence of a self-contained data interface compatible with the existing contact locations on these battery packs, including the lack of integral power conductors, today's thermistors cannot be used to provide a second level of safety by monitoring heat within either of these battery packs. All of the thermistors below are also limited in their ability to deliver their temperature information to an external device, such as an external battery charger, for example.
Temperature-Sensing Devices
Temperature-sensing devices, whether a positive temperature coefficient (PTC) or negative temperature coefficient (NTC) type, have not exhibited response times adequate to the rapid heat build-up associated with today's volatile battery chemistries. Nickel-Cadmium (NiCad) and Nickel-Metal-Hydride (NIMH) battery chemistries can show thermal runaway behavior during charge, but Lithium-Ion (Li-Ion) batteries are much more unstable. Li-Ion cells' sensitivity to charge overvoltages, and even to inappropriate trickle charging, can create sudden heat build-up in cells. If this temperature increase is not ameliorated promptly, the cell can explode.
Temperature-sensing is a reliable way to detect aberrant cell behavior. Early detection of even minor temperature increases inside a cell can prevent overcharging, thermal runaway, and the explosive consequences. Even cell venting, which is an accepted method of releasing a cell's internal pressure caused by beat, can be avoided by detecting temperature increases quickly.
Heat Affects Battery Efficiency and Life
The System Management Bus (SMBus) specifications (available from the standard body's web site: www.sbs-forum.org) commonly used to define safe laptop battery charging standards, specify not only an in-circuit thermistor, but also provide for an additional temperature sensor external to the battery's circuit board.
Temperature increases during battery cell charging have detrimental effects on the cell's chemistry, often with dangerous consequences. As one battery charger application engineer notes: "Temperature increases [within the cell] generate additional chemical reactions that are irreversible . . . For example, heat creates oxygen, which builds up pressure in a nickel-metal hydride (NiMH) cell. Not only is that an unsafe condition, but it reduces battery life, because it's nonreversible." Another application engineer indicates that: "As far as temperature is concerned, . . . there's a big difference between NiCad and NiMH. It's highly recommended to use a thermistor sensor as part of the primary or back-up [charge] termination method" (McKinnon, Cheryl, "Battery mission: to charge and to protect," Portable Design (October, 1997), pp. 33-43).
Even "smart" circuits in rechargeable batteries leave room for improvement. Traditional component-style thermistors are mounted on circuit boards located at one end of the battery-pack's plastic housing. There can be as many as 10 cells in a battery pack, yet a board-mounted thermistor can only be in relative proximity to the nearest cells at that end of the battery pack. Cells as much as 8 inches away from this board-mounted thermistor can overheat, and the remote temperature sensor at the opposite end of the battery pack will not indicate an over-heated condition for perhaps 15 seconds or more. This is an eternity when preventing a potential cell explosion.
Thermistor Response-Time Is a Function of Distance
Critical time-to-respond is determined by the distance between the heat source and the temperature-sensing mechanism. The model is analogous to the thermostat in a house. A household thermostat can only sense room temperature near to its location, so an over-heated room at the opposite end of the house is commonplace. Battery pack enclosures are just like a house. If there is a temperature sensor (thermostat) only at one end of the battery housing, detecting a distant over-heated cell (remote room) is impractical.
Typical Thermistor Response Times
Not only does distance between a heat source and a remote temperature sensing thermistor contribute to lack of adequate response time, but the inherent lag or delay in a thermistor design also increases total response time. Thin Film Detectors (TFDs), discussed below, can have a response lag time of 13.5-55 seconds when detecting the temperature of local air (as would be the case in a battery pack, where a single thermistor is sampling ambient air temperature within the enclosure). Average response times of PTC-style sensors are typically 20 seconds. Thus, placing such slow-responding temperature sensors at a distance from an over-heating cell only exacerbates problems with timely responses.
Remote Thermistors Don't Fit Existing Battery Pack Manufacture Configurations
Although SMBus specifications provide circuit board connections for a remote thermistor, this device is rarely implemented. This is a function of thermistor configurations. Temperature sensors manufactured by Keystone Thermometrics (St. Mary's, Pa.), Semitec (Babylon, N.Y.), among others, are configured as board-mountable components. These do not lend themselves to convenient placement in remote locations in a battery pack.
Some battery pack designers embed probe-type thermistors in the opening created by stacking cylindrical cells in a cluster of three. This approach allows for some improvement in thermistor response time by locating the temperature sensor adjacent to the cells. Mounting such devices, however, typically requires potting the probe. The potting compound has an insulator effect, thus degrading the efficacy of the thermistor. Also, few battery cavities accommodate a stacked cluster of three cells. In a cell phone or laptop computer, for example, cells are typically mounted side-by-side in a flat-pack configuration to minimize product thickness.
Cost is an Issue
Among the thinnest thermistors in the marketplace are thermocouples. Omega Corporation (Stamford, Conn.) fabricates thermocouples from Copper/Constantan. The thinnest bead available is 0.005 inches (for a maximum temperature of 400 degrees F. Although quite thin, these sensor probes are still very localized. Another prohibitive issue is that they are priced at about $17.00-25.00 each. If a thermocouple were attached to each cell in a nine-cell battery pack, the manufacturing cost would be $153.00-225.00 per battery pack. The retail price of such packs would be more than $500.00!
Semi-disposable rechargeable battery packs, commonly used in consumer-electronic products, require cost-effective temperature-sensing solutions. For example, more than 700,000 battery packs are manufactured each year for laptop computers. The board-mounted thermistor used today in such battery packs costs less than 30 cents.
Thin Film Detectors (TFDs) are flat-shaped platinum-resistance devices often used for temperature monitoring in wind tunnels or air conditioning systems. They measure a mere 0.250 inches and still can only sense a localized area. An Omega "Thin-Film Detector" costs $25 each, and only covers a surface area of 0.040.times.0.125 inches. The same company's thin-film RTD temperature sensor unit prices at $32-71. For a cellular phone's three-cell battery pack, for example, the cost of individual-cell temperature sensing would exceed $75, for a rechargeable battery product with a typical retail price of $50.
Thermal-Ribbon RTD Thermistors
Thermistors are available configured as "thermal-ribbon RTDs", such as the Minco (Minneapolis, Minn.) S17422. While conformably thin (0.5 mm) and flexible, these thermistors are limited by a requirement for two-conductor wires per sensor. Thus, a 10-cell battery pack would require 20 discrete wires running from 10 thermistors. These wire bundles would have to be run longitudinally in the "valleys" between round cells. Manufacturing such a battery pack would be complex and unnecessarily expensive.
With the advent of newer polymorphic cells that are not cylindrical in cross section, but are rectangular and flat, the complex wiring of thermal-ribbon RTD's would increase the overall size of a battery pack. Adding width or thickness to polymorphic packs defeats one of their primary advantages--a small cross-sectional profile.
Also, existing thermal-ribbon thermistors are fabricated with precious metals, such as platinum, which increases cost. To use six or more of these in a battery pack, so that each cell has its own temperature sensor in order to provide total temperature protection, adds considerably to the cost of a NiCad battery pack.
Polymeric PTC Materials
Raychem Corporation (Menlo Park, Calif.) manufactures Polyswitch Resettable Fuses that incorporate polymeric positive temperature coefficient (PTC) materials. The Polyswitch is used herein to illustrate the PTC class of thermally-conductive materials. Other manufacturers, such as DuPont (Cornwallis, N.C.), Keystone, and Omega have similar PTC chemistries, so by examining the Raychem PTC devices, all other similar products are assumed as equivalents. Composed of a matrix of crystalline organic polymer with dispersed conductive carbon particles, the quantity of conductive particles (which are carbon black) in the polymer matrix changes its physical properties to be less or more conductive. Precipitating this change of physical and electrical states are both the effects of temperature and electrical current, with current being the predominant force contributing to changes in resistivity.
Temperature inside a device, such as a battery pack, for example, is not a predictable means of triggering Raychem's PTC. The rate of heat loss within the battery pack must be less than the heat generated within a Polyswitch device. If the heat generated within a battery enclosure, more importantly heat at the location of the PTC, is greater than the heat loss of the polymeric PTC, the total energy required to make the Raychem device trigger (and thus stop an elevated temperature state within the cell) increases. The greater the heat transfer from the Polyswitch device to the environment, the slower the device's time-to-trip. This method of utilizing PTC polymer temperature sensing makes it extremely difficult to anticipate the change in the time it takes to trip the device. There is also a noticeable break or lag in the "time-to-trigger" with these types of devices, the time-lag created by the polymers' transition from an adiabatic to an non-adiabatic state.
Current-sensing, Not Temperature-sensing Drives Polyswitch Performance
A Polyswitch device may serve as an adequate fail-safe in overcurrent conditions during battery charging, since changes in electrical current-flow resulting from improper cell impedances favor triggered events caused by resistance changes in the Polyswitch device. Devices in this PTC class are not appropriate as early-warning sensors for detecting initial increases in individual cell temperature. Manufacturers of "smart" batteries, that have built-in circuitry to monitor cell voltage, current, and temperature, rely on Raychem's Polyswitch for over-current states. Polyswitch-type devices are not used for direct temperature monitoring because of their slow response times and trip-point lags.
The Role of Previous Printed-Ink Thermistors
Resistive printed-ink devices are commonly known. For example, U.S. Pat. No. 4,882,466 (Friel), is a two-electrode PTC that uses crystalline organic polymer and a conductive filler mixture. This material is applied to a substrate in the form of an ink. Friel's invention, as with the previously described Raychem PTC (the Friel patent was assigned to Raychem Corporation), reacts to electrical current applied to conductive ink to create a heater element. Heating (often referred to as "self-heating") is a generic trait of all positive temperature coefficient (PTC) thermistors. This self-heating characteristic is usually exploited, as it is in Friel, in devices that act as heaters. An example of this is the use of PTC devices as heaters for mirrors, where temperature response is employed to allow current to flow at or below a pre-defined temperature. Thus, an automobile's external side-view mirrors are heated only when the ambient temperature of the mirrored surface drops to a pre-specified temperature. Such self-heating characteristics only contribute to slower temperature-related response times when used inside a battery pack.
While Friel suggests that printable resistive ink traces can be applied to multi-layered substrates, the defined conductive paths are for heat generation through the conductance of electrical power. There is no provision for using conductors for data transmission, as is called for in the thermistor of the present invention. Also, Friel employs layers of conductive polymers and current-carrying electrodes which are attached or bonded in such a manner that they interact electrically, and are not insulated by appropriate dielectric materials.
Thermistor Surface Area Aids Response
Traditional thermistor size parameters have been driven by their applications. For laboratory, test equipment, and industrial applications, monitoring temperature at a well-defined location has required small form factors. Needle-thin probes, circular PTCs smaller than a dime, and rectangular postage-stamp-sized patches have been the norm. However, a single cylindrical 3/4-size battery cell has a surface area of approximately 5.25 inches. Present sensor design and form factors are not optimized for wide-area coverage.
For newer prismatic polymorphic cells, with typical dimensions of 1.times.4.5inches along their top and bottom surfaces, small surface-area thermistors are not well matched dimensionally.
Minco (Minneapolis, Minn.) manufactures a custom Thermal-Ribbon temperature sensor (Discoil) that can be ordered with dimensions as large as 2 feet square. While this sensor could cover one surface on an entire battery pack, it is somewhat thick (0.3 mm) and, more importantly, quite expensive. Discoil uses platinum elements, so it would be cost-prohibitive as a semi-disposable thermistor that covers an entire face of a battery pack enclosure. The dimensions of such a battery pack could easily exceed 24 square inches.
Ink-based PTCs Not Optimized
PTC devices, whether or not fabricated with conductive inks, are not optimized for battery temperature monitoring for several reasons. First, they are dependent on sizeable current flow (often 200 Ma or more) to create changes in temperature-response along the conductive ink surfaces. In theory, if one were to use such a PTC temperature-sensing device on a cell, it would only further warm the cell, as current from the battery charging circuit flowed through the sensor. This would be an inappropriate response, and diametrically opposed to the desired effect of keeping battery temperatures low.
Second, because PTC sensors rely on electrical current to achieve functionality, they create a drain on available power. Battery cells can overheat through rapid discharge, where the load imposed by a PTC thermistor only hastens the discharge rate. Even in a charge mode, a PTC thermistor on each cell in a 10-cell battery pack would contribute an excessive additional load, requiring extra power from the charging circuit.
Third, electrode resistivity is critical to the proper operation of PTC sensors. Friel describes specific formulaic electrode width-to-length ratios to establish a stable and repeatable temperature trigger. Electrodes are to be "as thin as possible," according to Friel, with average thicknesses (widths) of 0.0001 to 0.01 inches. Such precision requires close manufacturing tolerances. Furthermore, in an application in which the PTC unit is to be flexed, for example when wrapping a temperature sensor around a single battery cell, these thin traces can crack or delaminate from the substrate material.
Fourth, the lack of durability of such devices as Friel's and others discussed here, do not readily lend themselves to external attachment on pre-manufactured battery packs, where consumer wear-and-tear would degrade the thermistor's performance by constant handling. Because all cellular phones and laptops have removable battery packs, as do many other consumer electronic devices, the frailty of devices like Friel's creates durability concerns. Affixing PTC thermistor constructs to battery enclosures that are repeatedly inserted into, and removed from, battery bays or slots, indicate that abrasion would become an issue. Compounding the problem is the thickness of thermistors like Friel's and the thermal-ribbon and RTD configurations previously discussed. Especially troublesome is the thick mass created at the hardpoints where the thermistor's surface electrodes transition to (often thick) wire conductors. Battery packs usually fit snugly into molded cavities in their host devices, so thickness parameters must be extremely thin.
Fifth, the Friel device requires the inked electrodes to be printed on conductive surfaces onto which have been deposited or laminated resistive elements, such as electrodeposited copper or other conductive metals. Such a multiple-deposition or dual-layered process increases manufacturing complexity and unit cost.
Externally Attached Labels Using Thermally-Sensitive Materials
Attached labels on battery cells or battery packs are addressed by U.S. Pat. No. 5,626,978 (Weiss), wherein a label incorporates thermally-sensitive materials as part of a test circuit. Weiss does not allow for connectivity to a local or remote circuit, as would be the case with the present invention used with a "smart" battery's A/D circuitry. Weiss does not address any connectivity to external temperature monitoring or charger devices. Weiss' invention is restricted to detecting battery capacity, expressed as a visual display of a "fuel gauge." The information displayed on a battery label in certain modalities of the present invention is passive and unchanging, such as a company logo or user instructions. Weiss' invention, especially its display, is active. By depressing a selected area, Weiss' battery label displays remaining battery capacity. The present invention operates autonomously and without contacts or switches of any sort.
In summary, a variety of features and characteristics of today's thermistors leaves room for improvement in temperature monitoring of power sources, for example rechargeable and removable batteries (especially "smart" batteries), and external power supplies and battery chargers):
1). Traditional temperature-sensors, whether board-mountable electronic components, or attachable/placeable probes with wire leads, cannot be easily integrated into, or attached to, already-manufactured power sources, such as battery packs. PA1 2). Even if a battery pack, as an example of a power source, could be opened up, if the power source requiring temperature monitoring does not have an internal circuit board, traditional thermistors are not of much use. Lacking an internal circuit that provides A/D functions, as well as appropriate contacts on the battery housing, a power source's temperature data cannot be accessed by an external monitoring device. PA1 3). Power sources that do have internal A/D circuits and accessible data contacts on an external, cannot be readily modified to add additional temperature monitoring. New contact points on the power source-device's housing, dedicated to the new thermistor, would have to be created. These new contacts would have to be appropriately placed to interface with existing data contacts on the host device, for example, contacts in the battery cavity of a "smart" battery-equipped laptop computer would have to be modified to provide a data interface to the newly added thermistor. PA1 4). Slow response times, as a consequence of a thermistor's inherent lag, or "time-to-trigger." PA1 5). More importantly, slow response times due a thermistor's physical configuration create limitations as a remote temperature sensor. These physical characteristics include size, mounting or attachment requirements, and flexibility. PA1 6). The inability to attach or bond the thermistor to a cell or battery pack housing using materials that act as insulators. PA1 7). The need to have at least two conductors per thermistor. PA1 8). A PTC thermistor's inherent characteristics, dictated by materials, to self-heat from current flow, or to experience lag times due to changes in external and internal temperature variants. PA1 9). The use of power resources to energize current-flow-enabled "self-heating" PTCs that can further drain a battery's charge, or create additional load demands on a power supply or battery charger. PA1 10). Cost of thermistors, driven by fabrication materials primarily, in a consumer marketplace where low-cost battery packs are the norm. PA1 11). Small "foot print" or area coverage of thermistors. PA1 12). The need to redesign the battery pack enclosure, whether because of additional bulky wiring, or to accommodate cross-sectionally thick sensors. PA1 13). The requirement to hold close manufacturing tolerances to create ultra-thin strands of precious metals, or to control the complex dimensions of printed-ink electrodes. PA1 14). The need for specific substrates, which often must be coated with metals, and which are often costly to fabricate. PA1 15). Lack of durability, especially of thermistors which have a reasonably thin cross-sectional profile, when attached to the exterior of a removable battery pack. PA1 16). For an external application of a thermistor to an existing battery housing, the absence of an effective data interface to a "smart" battery, or the battery's host device.
Therefore, there exists a need for an improved apparatus for monitoring temperature that addresses one or more of the deficiencies in today's apparati for monitoring temperature.